May ,i^, ?9?^| 



kN^IjIHE 



2,»5 



saturated . and unsaturated hydrocaTbons of the 

 naphthienic of: polynuclear type, and, to a limited ex- 

 tent, of paraffins and aromatic hydrocarbons. 



The authors have made careful determinations of 

 viscosity by a modifietl form of Ostwald apparatus, 

 and discuss lubricating value in terms of this property 

 and chemical composition. The desideratum for a 

 lubricant appears to be low viscosity and good lubri- 

 cating power, as the lubricant is concerned in the 

 transformation of solid into liquid friction. On the 

 other hand, the mysterious property known as "body" 

 — that is, the power of maintaining a film (or film 

 strength) between shaft and bearings — is a vital factor. 

 This property of body seems to depend on surface 

 tension, and is not necessarily dependent on viscosity, 

 but appears to be connected with molecular weight. 

 The possible iso-colloid nature of heavy oils in explain- 

 ing high viscosity Is considered, and the view is put 

 forward that if this is the true nature of such oils, 

 the film strength may depend upon the relation of the 

 two phases present in the oil. 



The following are some of the authors' general con- 

 clusions : — Paraffinoid oils, though highly stable, have 

 little lubricating value, and the smaller the hydrogen 

 content, the greater the viscosity and durability ; it is 

 the polynuclear naphthenes, especially the unsaturated 

 components, which confer viscosity and film strength ; 

 the higher fractions of petroleum contain such un- 

 saturated members, which, when removed, exhibit 

 lower density, lower viscosity, higher molecular 

 weight, and, generally, lower lubricating value. 



It appears, then, that the true lubricant is an un- 

 saturated compound possessing the characteristic attri- 

 butes of such compounds, and this applies not only 

 to hydrocarbons, but also to fatty oils, such as rape, 

 castor, and olive oil ; in other words, unsaturation is a 

 feature of chemical activity which, in addition to 

 colour, taste, smell, and physiological properties, mani- 

 fests itself in lubricating value. Moreover, the col- 

 loidal condition of a good lubricant may also have to 

 be taken into consideration. 



The whole subject of liquid lubricants is an im- 

 portant one, and in the above brief summary the 

 authors have brought forward a contribution of a 

 highly suggestive character, which emphasises very 

 clearly the necessity for further comprehensive and 

 detailed study. 



STATISTICAL METHODS APPLIED TO 

 PRACTICAL PROBLEMS. 



A17E have received a copy of the presidential ad- 

 * * dress delivered before the Indian Science Con- 

 gress at Lahore in January last by Sir G. T. Walker, 

 Director-General of Observatories, India (Calcutta ; 

 Thacker, Spink, and Co.). The bulk of the address is 

 devoted to illustrating the application to important 

 practical problems of statistical methods. Illustra- 

 tions are given of the use of coefficients of correlation 

 in connection with crop prediction, and also to show 

 the extent of co-variation between the fever rate, 

 population, cost of rice, cultivated area, and rainfall, 

 so far as this can be determined by a first-degree 

 regression equation computed from the data of twenty- 

 five years. 



An interesting point is the apparently close associa- 

 tion between an increase of population and an increase 

 of fever. Sir G. T. Walker observes that this ""is so 

 directly opposed to widely accepted medical ideas that 

 for some months I did not treat it seriously. But on 

 my informing Major McKendrick, of Kasauli, of it he 

 urged me not to reject it lightly, and directed my atten- 

 tion tp some analysis of his which would suggest an 

 explanation. I have therefore examined the corre- 



NO. 2533, VOL, lOl] 



spopding relationship for all the ten provinces of India 

 for which sujfficierit data are available, and bave not 

 found a single negative coefficient. The average 'Value 

 is +05. For those of you who iare interested in sani- 

 tary matters 1 may briefly explain that the. s|Jread of 

 malaria among men depends upon the meeting of 

 healthy men with infected mosquitoes, and the spread 

 among mosquitoes on the encounters between infected 

 men and healthy mosquitoes ; hence the relations are 

 symmetrical from the mathematical, if not the aesthetic, 

 point of view, and an increase in the number of men 

 has essentially the same bad effect as an increase in 

 the number of mosquitoes." 



In view of the work which has been done upon the 

 method of variate difference correlation during recetit 

 years, it would be of interest further to explore the 

 association by the new method. Statisticians in this 

 country will appreciate the concluding passage of Sir 

 G. T. Walker's address: — "I hof>e that statistical 

 methods may before long be recognised as essential 

 for efficiency for the following reasons. First, a table 

 of data covering, say, fifty years gives any intelligent 

 man the same advantage as if he had carefully watched 

 the conditions for fifty years and had a perfect memory ; 

 secondly, employing a draughtsman to plot these data 

 will suggest relationships in a manner which would 

 otherwise require profound study of the figures; and 

 thirdly, employing a clerk to work out the correlation 

 coefficients and regression equation will give him with- 

 out effort trustworthy information about their relatipn- 

 ships which will distinguish direct from indirect, elfejcts, 

 and could be got in no other way.*' ,, ;>\ ,..",,,', 



METHODS^ ■ OF: GA:S WARFAJiE: 



'X'HE .issue' of the. Journal of. the Washington 

 -•• Academy of Sciences for February 4 last ii>cludes 

 a report, of: a. lecture ^y Prof. S,, J. M., .Auld, of the 

 British .MiHtary -Mission, on "Methods of Gas War- 

 fare," delivered before the Academy. Naturally in 

 the lecture, which is here summarised, attention is con- 

 fined to. a description of what the Germans have been 

 doing; nothing is said of the activities of the Entente 

 Powers. in' this direction. - ■ > 



The; first gas attack , was made by the Gefttlans in 

 April,; 1915, and the ; whole method of the wor was 

 changed. The attack was. made,, of course, against 

 men who were entirely, unprepared— absolutely unpro- 

 tected- 1 The, Germans expected no immediate if efalia- 

 tion, as they, had provided no protection for their own 

 men. A clear and unobstructed gap in the lines was 

 made, which was only closed by the Canadians, who 

 rallied on the left and advanced, in part through the 

 gas-cloud itself. 



The method first used by the Germans is simple, but 

 requires great preparation beforehand. A hole is dug 

 in the bottom of the trench close underneath the 

 parapet, and a gas cylinder is buried in this hole. It 

 is then covered first with a quilt of moss containing 

 potassium carbonate solution, and then with sand- 

 bags. When the attack is to be made the sand-bags 

 are taken off the cylinder, and each cylinder is con- 

 nected with a lead pipe which is bent over the top of 

 the parapet. A sand-bag is laid on the nozzle to 

 prevent the back "kick" of the outrushing gas from 

 throwing the pipe back into the trench. 



The attackers must know the direction and velocity 

 of the wind with certainty. Favourable conditions are 

 limited practically to wind velocities between twelve 

 and four miles an hour. A wind of more than twelve 

 miles an hour disperses the gas-cloud very rapidly. 

 An upward current of air is the worst- foe of gas. If 

 the trench line is very irregular it is likely tnat gas 

 will flow into a portion of one's own trenches. The 



